This is a follow up to #54 and bids-standard/bids-specification#627 (comment): The remove of "contributions" from contributors no longer associated with some "dynamic entity" (such as "steering committee") is somewhat not "in-line" with the idea of all-contributors which is "cumulative" - all categories a person contributed to some point in time. But it opens the way for establishing "collective names" (such as "Steering committee"), the exact listing of which then could be deduced given a range of commits in the repository with .all-contributors changing, and which might be needed for adequate reflection of participation for publications etc.
But may be such an approach should be applied to the rest of the contributors (especially if we start collecting data on issues/discussions), so then at each particular point in time we would get a list of contributors since previous release, and for "cumulative" just analyze all the "released states" to that point.
Alternative could be to keep timeline for each time of contribution type, thus ignoring the history which could be recovered from git. But IMHO it would be not only more tedious but also would grow the file with all the time periods (even if updated/populated only at release points...).
This is a follow up to #54 and bids-standard/bids-specification#627 (comment): The remove of "contributions" from contributors no longer associated with some "dynamic entity" (such as "steering committee") is somewhat not "in-line" with the idea of all-contributors which is "cumulative" - all categories a person contributed to some point in time. But it opens the way for establishing "collective names" (such as "Steering committee"), the exact listing of which then could be deduced given a range of commits in the repository with .all-contributors changing, and which might be needed for adequate reflection of participation for publications etc.
But may be such an approach should be applied to the rest of the contributors (especially if we start collecting data on issues/discussions), so then at each particular point in time we would get a list of contributors since previous release, and for "cumulative" just analyze all the "released states" to that point.
Alternative could be to keep timeline for each time of contribution type, thus ignoring the history which could be recovered from git. But IMHO it would be not only more tedious but also would grow the file with all the time periods (even if updated/populated only at release points...).